#Automatic generation of bilingual dictionaries based on semantic vector spaces

Dictionaries and phrase tables are the basis of modern statistical machine translation systems. Traditional methods for automatic generation of bilingual dictionaries depend on parallel bilingual corpora. Since parallel corpora is hard to acquire, newer methods only require comparable corpora in order to work. This paper develops a method that can automate the process of generating and extending dictionaries and phrase tables in an unsupervised way from comparable corpora.

##Requires python 2.7 and:

sudo apt-get install python python-dev

sudo pip install beautifulsoup4 sudo pip install numpy sudo pip install nltk sudo pip install goslate

TreeTagger and it's Python wrapper are also requiered- http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/

sudo python ~/tree-tagger/setup.py install